頁籤選單縮合
題名 | A Comparison of Neural Networks to Process Capability=神經網路在製程能力判定應用上之研究 |
---|---|
作者 | 陳榮方; 鄭文助; |
期刊 | 高雄工商專校學報 |
出版日期 | 19951200 |
卷期 | 25 1995.12[民84.12] |
頁次 | 頁183-190 |
分類號 | 494.56 |
語文 | eng |
關鍵詞 | 製程能力; 神經網路; 倒傳遞網路; Process capability; Artificial neural networks; Back propagation network; |
中文摘要 | 本研究係探討神經網路中倒傳遞網路在製程能力上的判定能力,用以和品質管理 上的傳統統計檢定方法比較。 本文採用 C �洎�為製程能力的指數,以檢定雙尾的製程狀況 。並以 3 層的倒傳遞神經網路模擬製程狀況。所有的訓練和測試資料均由 EXCEL 軟體模擬 而得,並用 NeuralWorks Explorer 神經網路軟體執行、分析後証明倒傳遞神經網路比傳統 的統計方法在對製程能力的檢定上更具有判斷能力。 |
英文摘要 | A comparison of the performance of a back propagation aritificial neural network to traditional process capability decision criteria of quality management is drawn using computer simulation experiments. Process capability index in this paper is used C �� to summarize a process's potential to meet two-sided specification. The neural network choose a three-layer back propagation network to simulate a manufature's process and to check it's potential. The data of training and testing set were generated by computer simulation with EXCEL. The result after runing NeuralWorks Explorer verify that back propagation is more excellent than traditional statistic techniqe to determine where a manufacture's process is stable or not. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。